Skip to main content

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index] [List Home]
Re: [4diac-dev] Migrating forte to +11

Hi Jose,



On 08.06.20 11:27, Jose Cabral wrote:
in the past weeks, forte code from the +11 standard were spotted in the core and some communication modules, giving problems to old compilers not supporting this standard. The core code in branch 1.12.x is now

It would be interesting to know which compilers are affected. I think it makes a difference if it's a vendor-specific compiler with no upgrade option or just older versions of well-known compilers.


I just want to open the discussion to know your opinions about if we should start moving forte’s code to the standard +11 officially from the next version or not.

Depending on the above question, I would go as far as moving straight to C++17. It is the last small upgrade in the C++11 era and the core features are well supported by the major compilers by now. Speaking of Linux, there is probably no distribution anymore that is still in its support life cycle and does not have a C++17 compiler.

Most importantly, C++17 would give us std::variant, which could simplify the ANY type a lot and make it more type safe, while keeping its efficiency.

C++20, on the other hand, is definitely too new. Coroutines might be an amazing way to simplify event queues and execution threads, but it will probably take a long time until this appears in compilers for more exotic platforms.

Regards,
--
Dipl.-Inform. Jörg Walter
Senior Researcher
R&D Group Distributed Computation and Communication

OFFIS e.V. - Institut für Informatik
FuE Bereich Produktion | R&D Division Manufacturing
Escherweg 2, 26121 Oldenburg - Germany
Phone/Fax.: +49 441 9722-729 / -282
E-Mail: joerg.walter@xxxxxxxx
URL: http://www.offis.de

Registergericht: Amtsgericht Oldenburg VR 1956
Vorstand: Prof. Dr.-Ing. Wolfgang H. Nebel (Vorsitzender),
    Prof. Dr. techn. Susanne Boll-Westermann, Prof. Dr.-Ing. Axel Hahn,
    Prof. Dr.-Ing. Andreas Hein, Prof. Dr. Sebastian Lehnhoff


Back to the top